Unraveling the prognostic heterogeneity in patients with advanced melanoma between Australia (OZ) and the United States (US): Preliminary report of the PHAMOUS study.

Abstract

8516 Background: The AJCC Melanoma Task Force analyzed nearly 10,000 patients (pts) with either index or newly diagnosed (dx) distant metastatic melanoma (MM) in order to update the 7th Edition Cancer Staging Manual. A significant disparity in overall survival (OS) from date of MM dx was observed in pts from OZ vs. the US. We hypothesized that this disparity related to either an imbalance in prognostic or treatment factors, and/or tumor biology. Methods: Pts dx with MM between 1/02 and 1/07 known to the MIA (OZ), BIDMC and MDACC (US) surgical units and seen by the associated medical oncology units were identified. Data were collected on prognostic factors analyzed for the AJCC (including more complete LDH levels) as well as additional diagnostic, clinical, and treatment related variables including: number of metastatic sites (met sites), ECOG performance status (PS), diagnostic methods, intensity of follow-up (f/up), extent of CNS disease, as well as treatment approaches. Pts were compared between OZ and US using the Kruskal-Wallis and chi-square tests as appropriate. Cox proportional hazards regression was used to identify independent predictors of OS. Results: Data were collected from 1,031 pts (313 OZ, 718 US). Age and gender at MM dx were similar between the two populations. A significant difference in OS from date of MM dx by country was again observed. The median OS in the US was 15.90 mos (95% CI 14.36, 17.97) compared to 10.18 mos in OZ (95% CI 8.97, 11.60) (p < 0.0001). In a multivariate analysis including site of stage IV disease (M1a vs. M1b/M1c), number of met sites (1 vs. > 1), serum LDH (normal vs. high) and ECOG PS (0 vs. ≥ 1), country remained a significant independent predictor of OS (HR 1.39, 95% CI 1.15-1.68, p = 0.0006). Conclusions: Median OS in pts with MM was significantly worse in OZ than in the US even when major established prognostic variables were included in the model. These data suggest that yet to be discovered clinical or tumor-related prognostic factors or biases related to differences in f/up intensity and/or treatment may explain this survival disparity. The potential independent contribution of these additional factors is currently being explored. No significant financial relationships to disclose.